Do Revenue and Margins Still Matter in AI?

Episode Summary

Executive Summary: David George argues that the rise of large private markets and AI has rewritten growth investing: great companies can stay private longer, winners are larger, and venture/growth funds can still generate strong returns by backing the highest-quality founders early and later. He emphasizes ROIC, market pull, strength-of-strengths, and adapting to new business models where AI drives faster scaling, new moats, and potentially major shifts in labor spend.

Main Topics: Why large growth funds can still outperform (Priority: 5/5): George defends larger fund sizes by pointing to historical outperformance, concentration in big winners, and the expansion of private markets that now house many of the biggest value-creation opportunities. Private markets vs public markets (Priority: 5/5): He argues that many companies that once would have gone public now remain private longer, changing liquidity dynamics but not necessarily competitive dynamics, while shifting more return creation into private markets. AI’s impact on company evaluation (Priority: 5/5): George explains how AI changes underwriting: higher emphasis on retention, engagement, customer pull, and productivity impact, while being more forgiving on gross margins if AI is actually being used. Founder quality and investing in strength of strengths (Priority: 5/5): He repeatedly stresses backing exceptional founders with obvious spikes, even when there are weaknesses, and avoiding the mistake of over-weighting hypothetical future competition. Competition, market pull, and kingmaking (Priority: 4/5): The discussion covers when capital can matter, when kingmaking is real or overrated, and why the best companies usually win because they already have product-market pull and preferential attachment. AI application winners and business-model shifts (Priority: 4/5): He sees the biggest disruption in business model, UI/workflow, and data access, with vertical AI companies like customer support, radiology, and operations software potentially creating large new incumbents. Personalized excitement: health and robotics (Priority: 3/5): George closes by highlighting personal health management and robotics as likely major AI categories over the next decade, both as consumer experiences and investment opportunities.

Key Arguments: Large venture/growth funds can still produce excellent multiples because the biggest private-market winners are large enough to drive fund returns. The private market has absorbed more of the value creation that used to occur in public markets, so historical asset-allocation models are outdated. Public/private status does not primarily determine competitive dynamics; it mostly changes liquidity and cost of capital. In AI, revenue quality depends less on time-based renewal history and more on retention, engagement, and evidence of real customer pull. Investing should prioritize founder strength-of-strengths rather than obsessing over weaknesses or fear of theoretical competition. Business-model shifts, workflow changes, and access to proprietary data are the main ways AI startups can beat incumbents. Gross margins can be lower in AI today if the product delivers real value and adoption; the key is whether AI is actually being used. Kingmaking is limited; capital helps strongest companies scale, but money alone rarely creates a winner from scratch. The best AI companies may be much larger than previous-generation software companies because they can create surplus value and reallocate labor spend to technology. The future of AI will likely include specialized winners in consumer, B2B, and infrastructure, rather than one model company eating everything.

Data Points: Best-performing fund size: $1 billion - George says the firm’s best-performing fund in history is a $1 billion fund. Databricks return: 7x the fund - Cited as an example of one large fund’s winners. Coinbase return: 5x the fund - Used to support the claim that large funds can perform strongly. Private market size: Over $5 trillion - George says the private market has grown substantially and now holds much of the opportunity set. Top IPO gain split: 47% seed-to-Series B; 53% Series C onward - Analysis of the 50 top IPOs from 2017–2025 showing where gains accrue. Public companies decline: Cut in half over 20 years - Used to argue that small-cap public opportunity has shrunk. Russell 2500 ROIC: 7.5% down to 3% - Illustrates deterioration in public small-cap quality. Top 10 companies in the world: 8 of 10 are U.S. West Coast tech companies - Supports the case for venture exposure. CH Robinson productivity gain: 40% increase in shipments per person per day - Example of AI-driven operational improvement. CH Robinson operating margin: Up 680 basis points - Attributed to effective AI implementation. Microsoft headcount: Down 6% over the last year or so - Mentioned as a sign of AI-related efficiency and labor shifts. Average renter spend on rent: 30% of disposable income - Used to justify the rental-innovation thesis behind Flow. Follow-on activity in growth fund: About half of investments - George says roughly 50% of growth fund activity is follow-ons from existing venture companies. Origination follow-ons from growth fund: Another 15% - Additional follow-ons from companies originally funded by the growth team. Fully net new companies: About a third - Remaining share of growth fund investments comes from new relationships.

Pivotal Quotes: "If you overweight the fear of future theoretical competition, you can always talk yourself out of making an investment." — David George: Explaining why he avoids over-indexing on hypothetical competitive threats when backing strong founders. "The number one way to measure a company is ultimately return on invested capital." — David George: His framework for evaluating company quality, especially in AI and growth investing. "Our best-performing fund in the history of the firm is actually a $1 billion fund." — David George: A direct rebuttal to the claim that large venture funds cannot generate strong returns.

Implications: For investors, the main lesson is to update frameworks for a world where private-market winners are bigger, AI adoption is faster, and quality is better measured by pull, usage, and ROIC than by old SaaS heuristics. For founders, category leadership and speed now matter even more.